MDEV-18464: Port kill_one_trx fixes from 10.4 to 10.1

Jan Lindström · Jan Lindström · commit 21b2fada7ab7 · 2019-03-28T07:40:03.000+02:00
Pushed the decision for innodb transaction and system
locking down to lock0lock.cc level. With this,
we can avoid releasing these mutexes for executions
where these mutexes were acquired upfront.

This patch will also fix BF aborting of native threads, e.g.
threads which have declared wsrep_on=OFF. Earlier, we have
used, for innodb trx locks, was_chosen_as_deadlock_victim
flag, for marking inodb transactions, which are victims for
wsrep BF abort. With native threads (wsrep_on==OFF), re-using
was_chosen_as_deadlock_victim flag may lead to inteference
with real deadlock, and to deal with this, the patch has added new
flag for marking wsrep BF aborts only: victim=true

Similar way if replication decides to abort one of the threads
we mark victim by: victim=true

innobase_kill_query
	Remove lock sys and trx mutex handling.

wsrep_innobase_kill_one_trx
	Mark victim trx with victim=true

trx0trx.h
	Remove trx_abort_t type and abort type variable from
	trx struct. Add victim variable to trx.

wsrep_kill_victim
	Remove abort_type

lock_report_waiters_to_mysql
	Take also trx mutex and mark trx as a victim for
	replication abort.

lock_trx_handle_wait_low
	New low level function to check whether the transaction
	has already been rolled back because it was selected as
	a deadlock victim, or if it has to wait then cancel
	the wait lock.

lock_trx_handle_wait
	If transaction is not marked as victim take lock sys
	and trx mutex before calling lock_trx_handle_wait_low
	and release them after that.

row_search_for_mysql
	Remove lock sys and trx mutex taking and releasing.

trx_rollback_to_savepoint_for_mysql_low
trx_commit_in_memory
	Clean up victim variable.
diff --git a/storage/innobase/handler/ha_innodb.cc b/storage/innobase/handler/ha_innodb.cc
@@ -4929,8 +4929,6 @@ static void innobase_kill_query(handlerton*, THD* thd, enum thd_kill_levels)
 		/* if victim has been signaled by BF thread and/or aborting
 		   is already progressing, following query aborting is not necessary
 		   any more.
-		   Also, BF thread should own trx mutex for the victim, which would
-		   conflict with trx_mutex_enter() below
 		*/
 		DBUG_VOID_RETURN;
 	}
@@ -4939,34 +4937,8 @@ static void innobase_kill_query(handlerton*, THD* thd, enum thd_kill_levels)
 	if (trx_t* trx = thd_to_trx(thd)) {
 		ut_ad(trx->mysql_thd == thd);
 
-		switch (trx->abort_type) {
-#ifdef WITH_WSREP
-		case TRX_WSREP_ABORT:
-			break;
-#endif
-		case TRX_SERVER_ABORT:
-			if (!wsrep_thd_is_BF(trx->mysql_thd, FALSE)) {
-				lock_mutex_enter();
-			}
-			/* fall through */
-		case TRX_REPLICATION_ABORT:
-			trx_mutex_enter(trx);
-		}
 		/* Cancel a pending lock request if there are any */
 		lock_trx_handle_wait(trx);
-		switch (trx->abort_type) {
-#ifdef WITH_WSREP
-		case TRX_WSREP_ABORT:
-			break;
-#endif
-		case TRX_SERVER_ABORT:
-			if (!wsrep_thd_is_BF(trx->mysql_thd, FALSE)) {
-				lock_mutex_exit();
-			}
-			/* fall through */
-		case TRX_REPLICATION_ABORT:
-			trx_mutex_exit(trx);
-		}
 	}
 
 	DBUG_VOID_RETURN;
@@ -18683,6 +18655,12 @@ wsrep_innobase_kill_one_trx(
 		    wsrep_thd_ws_handle(thd)->trx_id);
 
 	wsrep_thd_LOCK(thd);
+
+	/* We mark this as victim transaction, which is already marked
+	as BF victim. Both trx mutex and lock_sys mutex is held until
+	this victim has aborted. */
+	victim_trx->victim = true;
+
         DBUG_EXECUTE_IF("sync.wsrep_after_BF_victim_lock",
                  {
                    const char act[]=
@@ -18866,7 +18844,7 @@ wsrep_abort_transaction(
 	my_bool signal)
 {
 	DBUG_ENTER("wsrep_innobase_abort_thd");
-	
+
 	trx_t* victim_trx	= thd_to_trx(victim_thd);
 	trx_t* bf_trx		= (bf_thd) ? thd_to_trx(bf_thd) : NULL;
 
@@ -18878,12 +18856,10 @@ wsrep_abort_transaction(
 	if (victim_trx) {
 		lock_mutex_enter();
 		trx_mutex_enter(victim_trx);
-		victim_trx->abort_type = TRX_WSREP_ABORT;
 		int rcode = wsrep_innobase_kill_one_trx(bf_thd, bf_trx,
                                                         victim_trx, signal);
 		trx_mutex_exit(victim_trx);
 		lock_mutex_exit();
-		victim_trx->abort_type = TRX_SERVER_ABORT;
 		wsrep_srv_conc_cancel_wait(victim_trx);
 		DBUG_RETURN(rcode);
 	} else {
diff --git a/storage/innobase/include/trx0trx.h b/storage/innobase/include/trx0trx.h
@@ -1,7 +1,7 @@
 /*****************************************************************************
 
 Copyright (c) 1996, 2016, Oracle and/or its affiliates. All Rights Reserved.
-Copyright (c) 2015, 2018, MariaDB Corporation.
+Copyright (c) 2015, 2019, MariaDB Corporation.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
@@ -623,7 +623,6 @@ struct trx_lock_t {
 					lock_sys->mutex. Otherwise, this may
 					only be modified by the thread that is
 					serving the running transaction. */
-
 	mem_heap_t*	lock_heap;	/*!< memory heap for trx_locks;
 					protected by lock_sys->mutex */
 
@@ -695,14 +694,6 @@ lock_rec_convert_impl_to_expl()) will access transactions associated
 to other connections. The locks of transactions are protected by
 lock_sys->mutex and sometimes by trx->mutex. */
 
-enum trx_abort_t {
-	TRX_SERVER_ABORT = 0,
-#ifdef WITH_WSREP
-	TRX_WSREP_ABORT,
-#endif
-	TRX_REPLICATION_ABORT
-};
-
 struct trx_t{
 	ulint		magic_n;
 
@@ -880,8 +871,12 @@ struct trx_t{
 	/*------------------------------*/
 	THD*		mysql_thd;	/*!< MySQL thread handle corresponding
 					to this trx, or NULL */
-	trx_abort_t	abort_type;	/*!< Transaction abort type*/
-
+	bool		victim;		/*!< This transaction is
+					selected as victim for abort
+					either by replication or
+					high priority wsrep thread. This
+					field is protected by trx and
+					lock sys mutex. */
 	const char*	mysql_log_file_name;
 					/*!< if MySQL binlog is used, this field
 					contains a pointer to the latest file
diff --git a/storage/innobase/lock/lock0lock.cc b/storage/innobase/lock/lock0lock.cc
@@ -1,7 +1,7 @@
 /*****************************************************************************
 
 Copyright (c) 1996, 2016, Oracle and/or its affiliates. All Rights Reserved.
-Copyright (c) 2014, 2018, MariaDB Corporation.
+Copyright (c) 2014, 2019, MariaDB Corporation.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
@@ -1793,10 +1793,8 @@ wsrep_kill_victim(
 				}
 			}
 
-			lock->trx->abort_type = TRX_WSREP_ABORT;
 			wsrep_innobase_kill_one_trx(trx->mysql_thd,
 				(const trx_t*) trx, lock->trx, TRUE);
-			lock->trx->abort_type = TRX_SERVER_ABORT;
 		}
 	}
 }
@@ -4782,12 +4780,11 @@ lock_report_waiters_to_mysql(
 			if (w_trx->id != victim_trx_id) {
 				/* If thd_report_wait_for() decides to kill the
 				transaction, then we will get a call back into
-				innobase_kill_query. We mark this by setting
-				current_lock_mutex_owner, so we can avoid trying
-				to recursively take lock_sys->mutex. */
-				w_trx->abort_type = TRX_REPLICATION_ABORT;
+				innobase_kill_query.*/
+				trx_mutex_enter(w_trx);
+				w_trx->victim = true;
 				thd_report_wait_for(mysql_thd, w_trx->mysql_thd);
-				w_trx->abort_type = TRX_SERVER_ABORT;
+				trx_mutex_exit(w_trx);
 			}
 			++i;
 		}
@@ -7967,16 +7964,7 @@ lock_trx_release_locks(
 	lock_mutex_exit();
 }
 
-/*********************************************************************//**
-Check whether the transaction has already been rolled back because it
-was selected as a deadlock victim, or if it has to wait then cancel
-the wait lock.
-@return DB_DEADLOCK, DB_LOCK_WAIT or DB_SUCCESS */
-UNIV_INTERN
-dberr_t
-lock_trx_handle_wait(
-/*=================*/
-	trx_t*	trx)	/*!< in/out: trx lock state */
+inline dberr_t lock_trx_handle_wait_low(trx_t* trx)
 {
 	ut_ad(lock_mutex_own());
 	ut_ad(trx_mutex_own(trx));
@@ -7993,6 +7981,32 @@ lock_trx_handle_wait(
 	return DB_LOCK_WAIT;
 }
 
+/*********************************************************************//**
+Check whether the transaction has already been rolled back because it
+was selected as a deadlock victim, or if it has to wait then cancel
+the wait lock.
+@return DB_DEADLOCK, DB_LOCK_WAIT or DB_SUCCESS */
+UNIV_INTERN
+dberr_t
+lock_trx_handle_wait(
+/*=================*/
+	trx_t*	trx)	/*!< in/out: trx lock state */
+{
+	if (!trx->victim) {
+		lock_mutex_enter();
+		trx_mutex_enter(trx);
+	}
+
+	dberr_t err = lock_trx_handle_wait_low(trx);
+
+	if (!trx->victim) {
+		lock_mutex_exit();
+		trx_mutex_exit(trx);
+	}
+
+	return err;
+}
+
 /*********************************************************************//**
 Get the number of locks on a table.
 @return number of locks */
diff --git a/storage/innobase/row/row0sel.cc b/storage/innobase/row/row0sel.cc
@@ -4746,11 +4746,7 @@ row_search_for_mysql(
 			a deadlock and the transaction had to wait then
 			release the lock it is waiting on. */
 
-			lock_mutex_enter();
-			trx_mutex_enter(trx);
 			err = lock_trx_handle_wait(trx);
-			lock_mutex_exit();
-			trx_mutex_exit(trx);
 
 			switch (err) {
 			case DB_SUCCESS:
diff --git a/storage/innobase/trx/trx0roll.cc b/storage/innobase/trx/trx0roll.cc
@@ -1,7 +1,7 @@
 /*****************************************************************************
 
 Copyright (c) 1996, 2017, Oracle and/or its affiliates. All Rights Reserved.
-Copyright (c) 2016, 2018, MariaDB Corporation.
+Copyright (c) 2016, 2019, MariaDB Corporation.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
@@ -370,6 +370,7 @@ trx_rollback_to_savepoint_for_mysql_low(
 	trx_mark_sql_stat_end(trx);
 
 	trx->op_info = "";
+	trx->victim = false;
 
 	return(err);
 }
diff --git a/storage/innobase/trx/trx0trx.cc b/storage/innobase/trx/trx0trx.cc
@@ -1,7 +1,7 @@
 /*****************************************************************************
 
 Copyright (c) 1996, 2016, Oracle and/or its affiliates. All Rights Reserved.
-Copyright (c) 2015, 2018, MariaDB Corporation.
+Copyright (c) 2015, 2019, MariaDB Corporation.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
@@ -1339,11 +1339,7 @@ trx_commit_in_memory(
 	ut_ad(!trx->in_ro_trx_list);
 	ut_ad(!trx->in_rw_trx_list);
 
-#ifdef WITH_WSREP
-	if (trx->mysql_thd && wsrep_on(trx->mysql_thd)) {
-		trx->lock.was_chosen_as_deadlock_victim = FALSE;
-	}
-#endif
+	trx->victim = false;
 	trx->dict_operation = TRX_DICT_OP_NONE;
 
 	trx->error_state = DB_SUCCESS;
diff --git a/storage/xtradb/handler/ha_innodb.cc b/storage/xtradb/handler/ha_innodb.cc
@@ -5534,43 +5534,15 @@ static void innobase_kill_query(handlerton*, THD* thd, enum thd_kill_levels)
 		/* if victim has been signaled by BF thread and/or aborting
 		   is already progressing, following query aborting is not necessary
 		   any more.
-		   Also, BF thread should own trx mutex for the victim, which would
-		   conflict with trx_mutex_enter() below
 		*/
 		DBUG_VOID_RETURN;
 	}
 #endif /* WITH_WSREP */
 	if (trx_t* trx = thd_to_trx(thd)) {
 		ut_ad(trx->mysql_thd == thd);
 
-		switch (trx->abort_type) {
-#ifdef WITH_WSREP
-		case TRX_WSREP_ABORT:
-			break;
-#endif
-		case TRX_SERVER_ABORT:
-			if (!wsrep_thd_is_BF(trx->mysql_thd, FALSE)) {
-				lock_mutex_enter();
-			}
-			/* fall through */
-		case TRX_REPLICATION_ABORT:
-			trx_mutex_enter(trx);
-		}
 		/* Cancel a pending lock request if there are any */
 		lock_trx_handle_wait(trx);
-		switch (trx->abort_type) {
-#ifdef WITH_WSREP
-		case TRX_WSREP_ABORT:
-			break;
-#endif
-		case TRX_SERVER_ABORT:
-			if (!wsrep_thd_is_BF(trx->mysql_thd, FALSE)) {
-				lock_mutex_exit();
-			}
-			/* fall through */
-		case TRX_REPLICATION_ABORT:
-			trx_mutex_exit(trx);
-		}
 	}
 
 	DBUG_VOID_RETURN;
@@ -19723,6 +19695,12 @@ wsrep_innobase_kill_one_trx(
 		  (thd && wsrep_thd_query(thd)) ? wsrep_thd_query(thd) : "void");
 
 	wsrep_thd_LOCK(thd);
+
+	/* We mark this as victim transaction, which is already marked
+	as BF victim. Both trx mutex and lock_sys mutex is held until
+	this victim has aborted. */
+	victim_trx->victim = true;
+
         DBUG_EXECUTE_IF("sync.wsrep_after_BF_victim_lock",
                  {
                    const char act[]=
@@ -19911,12 +19889,10 @@ wsrep_abort_transaction(handlerton* hton, THD *bf_thd, THD *victim_thd,
 	if (victim_trx) {
 		lock_mutex_enter();
 		trx_mutex_enter(victim_trx);
-		victim_trx->abort_type = TRX_WSREP_ABORT;
 		int rcode = wsrep_innobase_kill_one_trx(bf_thd, bf_trx,
                                                         victim_trx, signal);
 		trx_mutex_exit(victim_trx);
 		lock_mutex_exit();
-		victim_trx->abort_type = TRX_SERVER_ABORT;
 		wsrep_srv_conc_cancel_wait(victim_trx);
 		DBUG_RETURN(rcode);
 	} else {
diff --git a/storage/xtradb/include/trx0trx.h b/storage/xtradb/include/trx0trx.h
@@ -1,7 +1,7 @@
 /*****************************************************************************
 
 Copyright (c) 1996, 2016, Oracle and/or its affiliates. All Rights Reserved.
-Copyright (c) 2015, 2018, MariaDB Corporation.
+Copyright (c) 2015, 2019, MariaDB Corporation.
 
 This program is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free Software
@@ -672,7 +672,6 @@ struct trx_lock_t {
 					lock_sys->mutex. Otherwise, this may
 					only be modified by the thread that is
 					serving the running transaction. */
-
 	mem_heap_t*	lock_heap;	/*!< memory heap for trx_locks;
 					protected by lock_sys->mutex */
 
@@ -744,14 +743,6 @@ lock_rec_convert_impl_to_expl()) will access transactions associated
 to other connections. The locks of transactions are protected by
 lock_sys->mutex and sometimes by trx->mutex. */
 
-enum trx_abort_t {
-	TRX_SERVER_ABORT = 0,
-#ifdef WITH_WSREP
-	TRX_WSREP_ABORT,
-#endif
-	TRX_REPLICATION_ABORT
-};
-
 struct trx_t{
 	ulint		magic_n;
 
@@ -930,8 +921,12 @@ struct trx_t{
 	/*------------------------------*/
 	THD*		mysql_thd;	/*!< MySQL thread handle corresponding
 					to this trx, or NULL */
-	trx_abort_t	abort_type;	/*!< Transaction abort type */
-
+	bool		victim;		/*!< This transaction is
+					selected as victim for abort
+					either by replication or
+					high priority wsrep thread. This
+					field is protected by trx and
+					lock sys mutex. */
 	const char*	mysql_log_file_name;
 					/*!< if MySQL binlog is used, this field
 					contains a pointer to the latest file
diff --git a/storage/xtradb/lock/lock0lock.cc b/storage/xtradb/lock/lock0lock.cc
diff --git a/storage/xtradb/row/row0sel.cc b/storage/xtradb/row/row0sel.cc
diff --git a/storage/xtradb/trx/trx0roll.cc b/storage/xtradb/trx/trx0roll.cc
diff --git a/storage/xtradb/trx/trx0trx.cc b/storage/xtradb/trx/trx0trx.cc
